1276
Low Blood-to-Cerebrospinal Fluid Passage of Sorbitol After Intravenous Infusion Roland Nau, DrMed; Torsten Dreyhaupt, DrMed; Herbert Kolenda, DrMed; and Hilmar W. Prange, DrMed Background and Purpose: Compared with mannitol, the osmotherapeutic agent sorbitol is less prone to accumulate in the blood and the same quantity may be infused in a smaller volume. Because of these advantageous characteristics, we studied the pharmacokinetics of sorbitol in serum and cerebrospinal fluid. Methods: Six patients (five women and one man; age range, 46-70 years) with an external ventriculostomy and suffering from brain edema due to cerebrovascular disease received sorbitol as part of their therapy. Before and after the first dose of 50 g infused over 20 minutes, sorbitol concentrations in serum and cerebrospinal fluid were determined repeatedly using an enzymatic procedure. Results: Maximal sorbitol concentrations ranged from 2,705 to 5,821 (median, 3,227) nig/1 in serum compared with 6.7-130.7 (median, 19.5) mg/1 in cerebrospinal fluid. Cerebrospinal fluid maxima were observed 0.17-3 hours after the end of the infusion. Sorbitol elimination in serum was adequately described by a two-compartment pharmacokinetic model (distribution half-life, 0.05-0.14 hour, elimination half-life, 0.23-0.61 hour). Elimination in cerebrospinal fluid followed a single-exponential decay and was considerably slower than that in serum (half-life, 1.3-7.7 hours). Conclusions: The maximal cerebrospinal fluid concentration/maximal serum concentration ratio was low for sorbitol, thus suggesting a small potential risk of inducing an increase of intracranial pressure after osmotherapy (rebound effect). (Stroke 1992;23:1276-1279) KEY WORDS • cerebrospinal fluid • pharmacokinetics • sorbitol
G
lycerol, mannitol, and sorbitol have been used since the early 1960s to reduce intracranial pressure (ICP).'- 3 While in the United States mannitol is most frequently used, in Europe the preferences differ from hospital to hospital. When considering the pharmacokinetic properties, each drug has characteristic advantages and disadvantages. Sorbitol can be infused intravenously at a concentration of 40% (wt:vol), so that large increases in serum osmolality can be achieved with small volumes of infusion. Because of its limited solubility in water, mannitol solutions are 20% at most, and glycerol infusions of more than 10% are contraindicated because they provoke hemolysis. In contrast to mannitol, which is eliminated only by the kidneys, sorbitol and glycerol are also subject to metabolism in various tissues, mainly the liver.4-6 For this reason the latter two drugs may be applied in renal insufficiency; however, they may affect blood levels of other carbohydrates. The passage of osmotically active drugs into the central nervous compartment is probably related to the so-called rebound effect, i.e., the increase of ICP after cessation of infusion of osmotic agents. Some authors From the Departments of Neurology (R.N., T.D., H.W.P.) and Neurosurgery (H.K.), University of Gottingen, Gottingen, FRG. Supported in part by a grant of Braun, Melsungen, FRG. Address for correspondence: Dr. med. Roland Nau, Department of Neurology, University of Gottingen, Robert-Koch-StraBe 40, D-3400 Gottingen, FRG. Received December 4, 1991; final revision received April 17, 1992; accepted April 21, 1992.
assert that the hazard of inducing a rebound effect is smaller with sorbitol than with mannitol.7 The present study was performed to investigate the pharmacokinetics of intravenous sorbitol in blood and cerebrospinal fluid (CSF) at a dose applied in osmotherapy. The findings are discussed in view of data available for mannitol8 and glycerol.9 Subjects and Methods Six patients (five women and one man; age range, 46-70 years) suffering from cerebral edema due to cerebrovascular accidents received 125 ml of a 40% solution of sorbitol (Tutofusin S40, Pfrimmer, Erlangen, FRG, or Sorbit 40, Braun, Melsungen, FRG) over 20 minutes as part of their therapy. The patients had undergone external ventriculostomy for occlusive hydrocephalus. In cases 1 and 2 hydrocephalus developed after compression of the fourth ventricle; in cases 4-6 blood coagula obstructed the aqueductus cerebri. For further details on the patients studied see Table 1. Arterial blood and ventricular CSF samples were drawn from indwelling catheters before beginning the sorbitol infusion to determine basal endogenous concentrations and at the end and 0.17,0.5,1, 2,3, 5, 8, and 11 hours after termination of the infusion. Blood and CSF were centrifuged for 10 minutes at 3,000g immediately after sampling. The supernatant was stored at -70°C until assaying, which was performed within 2 weeks. The study protocol was approved by the Ethics Committee of the University Hospital. Informed con-
Downloaded from http://stroke.ahajournals.org/ by guest on June 24, 2016
Nau et al Sorbitol Passage Into CSF
1277
TABLE 1. Characterization of Patients in Order of Increasing Maximal Sorbitol Concentration in CSF After Intravenous Infusion of 50 g Over 20 Minutes
Pt/age/sex
Body wt (kg)
Disease
Stroke to infusion (days)
CSF protein (mg/1)
CSF/serum albumin ratio (xlO" 3 )
CSF WBC/mm3
CSF
RBC/mm3
1/70/F
65
Pontine hemorrhage
5
99
1.9
6
43
2/46/F
65
Cerebellar infarction
5
339
5.7
13
5,888
3/46/F
60
Subarachnoid hemorrhage
3
182
2.8
1
1,669
4/48/F
70
Intracerebral hemorrhage
23
764
20.1
14
7,381
5/61/M
60
Intracerebral hemorrhage, hepatitis B
10
788
7.3
1
853
6/64/F
60
Intracerebral hemorrhage
4
1,734
27.4
5
1,195
CSF, cerebrospinal fluid; WBC, white blood cell count; RBC, red blood cell count; F, female; M, male.
sent to participation was obtained from either the patient or nearest relative. Sorbitol concentrations in serum and CSF were determined by means of an enzymatic assay based on the conversion of sorbitol to fructose by sorbitol dehydrogenase. This process was coupled to auxiliary enzymatic reactions leading to the formation of formazan with maximal absorption at 492 nm.10 The test kit was purchased from Boehringer Mannheim, Mannheim, FRG. Assays using the sorbitol dehydrogenase reaction have been previously validated in biological fluids611 (manufacturer's information). CSF and 1:3 diluted serum were deproteinized by heat (95°C for 5 minutes) and measured in appropriate dilutions (0.002-0.4 ml). At sample concentrations of 100 mg/1 the recovery of sorbitol as determined by an external standard was 100.1% and the interday coefficient of variation was 4.7% (n = 18). At 10 mg/1 recovery was 103.8% and the interday coefficient of variation 11.7% (n = 10). The detection limit was 0.2 mg/1. At -70°C sorbitol was stable in serum for at least 2 months; at room temperature no decline in sorbitol concentrations was seen during the first 20 hours. Xylitol, which is present in concentrations of 0.99). The use of triexponential equations did not improve data curve fitting. Owing to the slower elimination of sorbitol in CSF than in serum, the concentration-time curve in CSF lagged behind that in serum. The corresponding ratios between sorbitol concentrations in CSF and serum (CCSF/CS) were not constant. Dependent on time, CCSF/CS varied between 0.0016 (case 5 at the end of sorbitol infusion) and 176 (case 6 11 hours after the end of sorbitol administration). The AUCCSF/AUC S ratio as a measure of overall sorbitol passage into CSF ranged from 0.019 to 0.461 (median, 0.037). CmaxCsF, CmaxCsF/Cn,axs, and AUCCSF/AUCS correlated positively with the CSF protein content and the CSF/ serum albumin ratio (QAII,) (Table 1). The correlations between CraaxCSF and CSF protein content (Spearman's rank order correlation coefficient, r s =0.94), between CmaxCsF and QMb (rs=0.89), between AUCCSF/AUC S and QAU> (r s =0.83) and between CmaxCsF/CmiD,s and QAlb (rs=0.90) were all significant (p 1,000 red blood cells/ml in their CSF. Blood in CSF acts as a foreign body and, therefore, may have been involved in the development of an impaired blood-CSF barrier. However, because the interval from the hemorrhage to sorbitol infusion was at least 3 days, blood contamination should not have affected the results. Consequently, the AUCCSF/AUC S ratios did not correlate with the number of red blood cells in CSF (r s =-0.03). Elimination of sorbitol from CSF was several times slower than that in serum, leading to CSF concentrations that increased the CSF-to-serum concentration gradient in all patients during the elimination phase (Table 2). Because no additional barrier exists between CSF and the brain, the persistence of sorbitol in CSF probably reflected similar conditions in the interstitial fluid of the brain. The same phenomenon has been described in patients after intravenous glycerol treatment9 and probably occurs with mannitol as well, as can be deduced from results reported by Anderson and
Downloaded from http://stroke.ahajournals.org/ by guest on June 24, 2016
Nau et al Sorbitol Passage Into CSF coworkers.8 Unfortunately, the latter authors stopped measuring 4 hours after a 15-minute intravenous mannitol infusion. At that time mean plasma concentrations had fallen from approximately 4 g/1 to
Low Blood-to-Cerebrospinal Fluid Passage of Sorbitol After Intravenous Infusion Roland Nau, DrMed; Torsten Dreyhaupt, DrMed; Herbert Kolenda, DrMed; and Hilmar W. Prange, DrMed Background and Purpose: Compared with mannitol, the osmotherapeutic agent sorbitol is less prone to accumulate in the blood and the same quantity may be infused in a smaller volume. Because of these advantageous characteristics, we studied the pharmacokinetics of sorbitol in serum and cerebrospinal fluid. Methods: Six patients (five women and one man; age range, 46-70 years) with an external ventriculostomy and suffering from brain edema due to cerebrovascular disease received sorbitol as part of their therapy. Before and after the first dose of 50 g infused over 20 minutes, sorbitol concentrations in serum and cerebrospinal fluid were determined repeatedly using an enzymatic procedure. Results: Maximal sorbitol concentrations ranged from 2,705 to 5,821 (median, 3,227) nig/1 in serum compared with 6.7-130.7 (median, 19.5) mg/1 in cerebrospinal fluid. Cerebrospinal fluid maxima were observed 0.17-3 hours after the end of the infusion. Sorbitol elimination in serum was adequately described by a two-compartment pharmacokinetic model (distribution half-life, 0.05-0.14 hour, elimination half-life, 0.23-0.61 hour). Elimination in cerebrospinal fluid followed a single-exponential decay and was considerably slower than that in serum (half-life, 1.3-7.7 hours). Conclusions: The maximal cerebrospinal fluid concentration/maximal serum concentration ratio was low for sorbitol, thus suggesting a small potential risk of inducing an increase of intracranial pressure after osmotherapy (rebound effect). (Stroke 1992;23:1276-1279) KEY WORDS • cerebrospinal fluid • pharmacokinetics • sorbitol
G
lycerol, mannitol, and sorbitol have been used since the early 1960s to reduce intracranial pressure (ICP).'- 3 While in the United States mannitol is most frequently used, in Europe the preferences differ from hospital to hospital. When considering the pharmacokinetic properties, each drug has characteristic advantages and disadvantages. Sorbitol can be infused intravenously at a concentration of 40% (wt:vol), so that large increases in serum osmolality can be achieved with small volumes of infusion. Because of its limited solubility in water, mannitol solutions are 20% at most, and glycerol infusions of more than 10% are contraindicated because they provoke hemolysis. In contrast to mannitol, which is eliminated only by the kidneys, sorbitol and glycerol are also subject to metabolism in various tissues, mainly the liver.4-6 For this reason the latter two drugs may be applied in renal insufficiency; however, they may affect blood levels of other carbohydrates. The passage of osmotically active drugs into the central nervous compartment is probably related to the so-called rebound effect, i.e., the increase of ICP after cessation of infusion of osmotic agents. Some authors From the Departments of Neurology (R.N., T.D., H.W.P.) and Neurosurgery (H.K.), University of Gottingen, Gottingen, FRG. Supported in part by a grant of Braun, Melsungen, FRG. Address for correspondence: Dr. med. Roland Nau, Department of Neurology, University of Gottingen, Robert-Koch-StraBe 40, D-3400 Gottingen, FRG. Received December 4, 1991; final revision received April 17, 1992; accepted April 21, 1992.
assert that the hazard of inducing a rebound effect is smaller with sorbitol than with mannitol.7 The present study was performed to investigate the pharmacokinetics of intravenous sorbitol in blood and cerebrospinal fluid (CSF) at a dose applied in osmotherapy. The findings are discussed in view of data available for mannitol8 and glycerol.9 Subjects and Methods Six patients (five women and one man; age range, 46-70 years) suffering from cerebral edema due to cerebrovascular accidents received 125 ml of a 40% solution of sorbitol (Tutofusin S40, Pfrimmer, Erlangen, FRG, or Sorbit 40, Braun, Melsungen, FRG) over 20 minutes as part of their therapy. The patients had undergone external ventriculostomy for occlusive hydrocephalus. In cases 1 and 2 hydrocephalus developed after compression of the fourth ventricle; in cases 4-6 blood coagula obstructed the aqueductus cerebri. For further details on the patients studied see Table 1. Arterial blood and ventricular CSF samples were drawn from indwelling catheters before beginning the sorbitol infusion to determine basal endogenous concentrations and at the end and 0.17,0.5,1, 2,3, 5, 8, and 11 hours after termination of the infusion. Blood and CSF were centrifuged for 10 minutes at 3,000g immediately after sampling. The supernatant was stored at -70°C until assaying, which was performed within 2 weeks. The study protocol was approved by the Ethics Committee of the University Hospital. Informed con-
Downloaded from http://stroke.ahajournals.org/ by guest on June 24, 2016
Nau et al Sorbitol Passage Into CSF
1277
TABLE 1. Characterization of Patients in Order of Increasing Maximal Sorbitol Concentration in CSF After Intravenous Infusion of 50 g Over 20 Minutes
Pt/age/sex
Body wt (kg)
Disease
Stroke to infusion (days)
CSF protein (mg/1)
CSF/serum albumin ratio (xlO" 3 )
CSF WBC/mm3
CSF
RBC/mm3
1/70/F
65
Pontine hemorrhage
5
99
1.9
6
43
2/46/F
65
Cerebellar infarction
5
339
5.7
13
5,888
3/46/F
60
Subarachnoid hemorrhage
3
182
2.8
1
1,669
4/48/F
70
Intracerebral hemorrhage
23
764
20.1
14
7,381
5/61/M
60
Intracerebral hemorrhage, hepatitis B
10
788
7.3
1
853
6/64/F
60
Intracerebral hemorrhage
4
1,734
27.4
5
1,195
CSF, cerebrospinal fluid; WBC, white blood cell count; RBC, red blood cell count; F, female; M, male.
sent to participation was obtained from either the patient or nearest relative. Sorbitol concentrations in serum and CSF were determined by means of an enzymatic assay based on the conversion of sorbitol to fructose by sorbitol dehydrogenase. This process was coupled to auxiliary enzymatic reactions leading to the formation of formazan with maximal absorption at 492 nm.10 The test kit was purchased from Boehringer Mannheim, Mannheim, FRG. Assays using the sorbitol dehydrogenase reaction have been previously validated in biological fluids611 (manufacturer's information). CSF and 1:3 diluted serum were deproteinized by heat (95°C for 5 minutes) and measured in appropriate dilutions (0.002-0.4 ml). At sample concentrations of 100 mg/1 the recovery of sorbitol as determined by an external standard was 100.1% and the interday coefficient of variation was 4.7% (n = 18). At 10 mg/1 recovery was 103.8% and the interday coefficient of variation 11.7% (n = 10). The detection limit was 0.2 mg/1. At -70°C sorbitol was stable in serum for at least 2 months; at room temperature no decline in sorbitol concentrations was seen during the first 20 hours. Xylitol, which is present in concentrations of 0.99). The use of triexponential equations did not improve data curve fitting. Owing to the slower elimination of sorbitol in CSF than in serum, the concentration-time curve in CSF lagged behind that in serum. The corresponding ratios between sorbitol concentrations in CSF and serum (CCSF/CS) were not constant. Dependent on time, CCSF/CS varied between 0.0016 (case 5 at the end of sorbitol infusion) and 176 (case 6 11 hours after the end of sorbitol administration). The AUCCSF/AUC S ratio as a measure of overall sorbitol passage into CSF ranged from 0.019 to 0.461 (median, 0.037). CmaxCsF, CmaxCsF/Cn,axs, and AUCCSF/AUCS correlated positively with the CSF protein content and the CSF/ serum albumin ratio (QAII,) (Table 1). The correlations between CraaxCSF and CSF protein content (Spearman's rank order correlation coefficient, r s =0.94), between CmaxCsF and QMb (rs=0.89), between AUCCSF/AUC S and QAU> (r s =0.83) and between CmaxCsF/CmiD,s and QAlb (rs=0.90) were all significant (p 1,000 red blood cells/ml in their CSF. Blood in CSF acts as a foreign body and, therefore, may have been involved in the development of an impaired blood-CSF barrier. However, because the interval from the hemorrhage to sorbitol infusion was at least 3 days, blood contamination should not have affected the results. Consequently, the AUCCSF/AUC S ratios did not correlate with the number of red blood cells in CSF (r s =-0.03). Elimination of sorbitol from CSF was several times slower than that in serum, leading to CSF concentrations that increased the CSF-to-serum concentration gradient in all patients during the elimination phase (Table 2). Because no additional barrier exists between CSF and the brain, the persistence of sorbitol in CSF probably reflected similar conditions in the interstitial fluid of the brain. The same phenomenon has been described in patients after intravenous glycerol treatment9 and probably occurs with mannitol as well, as can be deduced from results reported by Anderson and
Downloaded from http://stroke.ahajournals.org/ by guest on June 24, 2016
Nau et al Sorbitol Passage Into CSF coworkers.8 Unfortunately, the latter authors stopped measuring 4 hours after a 15-minute intravenous mannitol infusion. At that time mean plasma concentrations had fallen from approximately 4 g/1 to
